Channel Estimation in Adaptive Modulation Systems

1. A method for an adaptive modulation communication system, comprising: receiving, over a communication channel, data encoded with a forward error correction (FEC) code; measuring first error information of said encoded data that is available before FEC decoding of said received encoded data; measuring second error information that is made available by said FEC decoding of said encoded data; and controlling transmission over said communication channel based on both said first error information and said second error information by providing a weighted combination of the first error information and the second error information for determining a transmission mode over said communication channel, wherein weights of the weighted combination depend on a rate of variation of a communication channel condition, and wherein a greater weight is given to said first error information than to said second error information in case the variation of the communication channel condition is faster than a predefined rate.

2. The method of claim 1 , wherein said FEC decoding of said encoded data is preceded by demodulating a modulated data signal received over the communication channel, wherein said FEC decoding consists in subsequently applying the FEC to the demodulated data signal, the first error information being the error information of the encoded data available before said demodulating.

3. The method of claim 2 , wherein the modulated data signal is a Quadrature Amplitude Modulation (QAM) data signal and said demodulating comprises QAM demapping symbols contained in the modulated data signal.

4. The method of claim 1 , wherein the first error information is a mean squared error level of the received encoded data.

5. The method of claim 2 , wherein the first error information is an estimate of a signal-to-noise plus interference ratio (SNIR) of the encoded data before said demodulating.

6. The method of claim 1 , wherein the second error information comprises a temporal average of a number of errors detected during said FEC decoding.

7. The method of claim 6 , wherein the temporal average is calculated over an integration time, wherein the integration time is less than the duration of a radio frame in said communication channel.

8. The method of claim 1 , further comprising the step of using said first error information for detecting variations of the communication channel condition faster than a predefined rate.

9. The method of claim 1 , wherein the first error information is compared with thresholds corresponding to modulation shift levels.

10. The method of claim 1 , wherein the FEC decoding includes single parity check (SPC) decoding and Reed-Solomon decoding, the second error information being the number of block errors detected by the SPC decoding.

11. The method of claim 3 , wherein measuring the first error information comprises integrating a SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, the method further comprising integrating a number of errors detected by SPC decoding over a period of half the radio frame, and wherein if the SNIR is decreasing and its rate of change exceeds the fast reaction threshold, the weight of the SNIR value is set to 1 and the weight of the detected errors from SPC decoding is set to 0; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold, the weight of the SNIR value is set to 0.5 and the weight of the detected errors from SPC decoding is set to 0.5; and wherein if the SNIR rate of change is below the standard threshold, the weight of the SNIR value is set to 0 and the weight of the detected errors from SPC decoding is set to 1.

12. The method of claim 3 , wherein measuring the first error information comprises integrating a SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, the method further comprising integrating a number of errors detected by SPC decoding over a period of half the radio frame, and wherein if the SNIR is decreasing and its rate of change and the detected errors rate of change exceed the fast reaction threshold, the weight of the SNIR value is set to 0.75 and the weight of the detected errors from SPC decoding is set to 0.25; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold and the detected errors rate of change exceeds the fast reaction threshold, the weight of the SNIR value is set to 0 and the weight of the detected errors average is set to 1; and wherein, regardless of the direction of change of the SNIR, if both the SNIR rate of change and the detected errors rate of change are between the standard and fast reaction thresholds, the weight of the SNIR is set to 0.25 and the weight of the detected errors average is set to 0.75; and wherein, regardless of the direction of change of the SNIR, if the SNIR rate of change is below the standard threshold, the weight of the SNIR is set to 0 and the weight of the detected errors average is set to 1.

13. The method of claim 1 , wherein the FEC decoding includes Low Density Parity Check iterative decoding.

14. The method of claim 13 , wherein the second error information is the number of failed parity checks at the first decoding iteration.

15. The method of claim 13 , wherein the second error information is the number of failed parity checks at a decoding iteration different from the first iteration.

16. The method of claim 3 , wherein measuring the first error information comprises integrating a SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, and the number of failed parity checks computed at the first decoding iteration is averaged over a number of LDPC blocks corresponding to a period of half the radio frame, wherein if the SNIR is decreasing and its rate of change exceeds the fast reaction threshold, the weight of the SNIR value is set to 1 and the weight of the averaged number of failed parity checks is set to 0; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold and the rate of change of the averaged number of failed parity checks exceeds the fast reaction threshold, the weight of the averaged SNIR value is set to 0 and the weight of the averaged number of failed parity checks is set to 1; and wherein, regardless of the direction of change of the SNIR, if both the SNIR rate of change and the rate of change of the averaged number of failed parity checks are between the standard and fast reaction thresholds, the weight of the SNIR is set to 0.25 and the weight of the averaged number of failed parity checks is set to 0.75; and wherein, regardless of the direction of change of the SNIR, if the SNIR rate of change is below the standard threshold, the weight of the SNIR is set to 0 and the weight of the averaged number of failed parity checks is set to 1.

17. The method of claim 3 , wherein measuring the first error information comprises integrating the SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, and the number of failed parity checks computed at said decoding iteration different from the first one is averaged over a number of LDPC blocks corresponding to a period of half the radio frame, wherein if the SNIR is decreasing and both its rate of change and the rate of change of the averaged number of failed parity checks exceed the fast reaction threshold, the weight of the SNIR value is set to 0.50 and the weight of the averaged number of failed parity checks is set to 0.50; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold and the rate of change of the averaged number of failed parity checks exceeds the fast reaction threshold, the weight of the averaged SNIR value is set to 0 and the weight of the averaged number of failed parity checks is set to 1; and wherein, regardless of the direction of change of the SNIR, if both the SNIR rate of change and the rate of change of the averaged number of failed parity checks are between the standard and fast reaction thresholds, the weight of the SNIR is set to 0.25 and the weight of the averaged number of failed parity checks is set to 0.75; and wherein, regardless of the direction of change of the SNIR, if the SNIR rate of change is below the standard threshold, the weight of the SNIR is set to 0 and the weight of the averaged number of failed parity checks is set to 1.

18. A device for an adaptive modulation communication system, comprising: an input device adapted to receive, from a communication channel, data encoded with a forward error correction (FEC) code; a FEC decoder connected downstream of the input device, for FEC decoding the received encoded data and providing error information determined by the FEC decoding; one or more measuring devices configured to: measure first error information of said encoded data before FEC decoding said received encoded data; measure second error information determined by said FEC decoder; a channel estimator configured to estimate a condition of the communication channel based on both said first error information and said second error information by providing a weighted combination of the first error information and the second error information for determining a transmission mode over said communication channel, wherein weights of the weighted combination depend on a rate of variation of a communication channel condition, and wherein a greater weight is given to said first error information than to said second error information in case the variation of the communication channel condition is faster than a predefined rate.

19. The device of claim 18 , further comprising a demodulator connected between the input device and the FEC decoder, the one or more measuring devices being adapted to measure the first error information of the encoded data between the demodulator and the input device.

20. The device of claim 19 , wherein the demodulator is a Quadrature Amplitude Modulation demodulator.

21. The device of claim 18 , wherein the first error information is a Mean Squared Error level of the received encoded data.

22. The device of claim 21 , wherein the first error information is an estimate of a signal-to-noise plus interference ratio, SNIR, of the encoded data upstream of said demodulator.

23. The device of claim 18 , wherein the second error information comprises a temporal average of a number of errors detected by said FEC decoder during said FEC decoding.

24. The device of claim 23 , wherein the temporal average is calculated over an integration time which is less than the duration of a radio frame in said communication channel.

25. The device of claim 20 , wherein measuring the first error information comprises integrating a SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, the method further comprising integrating a number of errors detected by SPC decoding over a period of half the radio frame, and wherein if the SNIR is decreasing and its rate of change exceeds the fast reaction threshold, the weight of the SNIR value is set to 1 and the weight of the detected errors from SPC decoding is set to 0; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold, the weight of the SNIR value is set to 0.5 and the weight of the detected errors from SPC decoding is set to 0.5; and wherein if the SNIR rate of change is below the standard threshold, the weight of the SNIR value is set to 0 and the weight of the detected errors from SPC decoding is set to 1.

26. The device of claim 22 , wherein measuring the first error information comprises integrating SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, the method further comprising integrating a number of errors detected by SPC decoding over a period of half the radio frame, and wherein if the SNIR is decreasing and its rate of change and the detected errors rate of change exceed the fast reaction threshold, the weight of the SNIR value is set to 0.75 and the weight of the detected errors from SPC decoding is set to 0.25; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold and the detected errors rate of change exceeds the fast reaction threshold, the weight of the SNIR value is set to 0 and the weight of the detected errors average is set to 1; and wherein, regardless of the direction of change of the SNIR, if both the SNIR rate of change and the detected errors rate of change are between the standard and fast reaction thresholds, the weight of the SNIR is set to 0.25 and the weight of the detected errors average is set to 0.75; and wherein, regardless of the direction of change of the SNIR, if the SNIR rate of change is below the standard threshold, the weight of the SNIR is set to 0 and the weight of the detected errors average is set to 1.

27. The device of claim 18 , wherein the channel estimator is adapted to compare the first error information with thresholds corresponding to modulation shift levels.

28. The device of claim 18 , wherein the FEC decoder includes a Single Parity Check decoder and Reed-Solomon decoder, the second error information being the number of block errors detected by the Single Parity Check decoder.

29. The device of claim 18 , wherein the FEC decoder includes a Low Density Parity Check iterative decoder.

30. The device of claim 29 , wherein the second error information is the number of failed parity checks at the first decoding iteration.

31. The device of claim 29 , wherein the second error information is the number of failed parity checks at a decoding iteration different from the first decoding iteration.

32. The device of claim 20 , wherein measuring the first error information comprises integrating a SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, and the number of failed parity checks computed at the first decoding iteration is averaged over a number of LDPC blocks corresponding to a period of half the radio frame, wherein if the SNIR is decreasing and its rate of change exceeds the fast reaction threshold, the weight of the SNIR value is set to 1 and the weight of the averaged number of failed parity checks is set to 0; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold and the rate of change of the averaged number of failed parity checks exceeds the fast reaction threshold, the weight of the averaged SNIR value is set to 0 and the weight of the averaged number of failed parity checks is set to 1; and wherein, regardless of the direction of change of the SNIR, if both the SNIR rate of change and the rate of change of the averaged number of failed parity checks are between the standard and fast reaction thresholds, the weight of the SNIR is set to 0.25 and the weight of the averaged number of failed parity checks is set to 0.75; and wherein, regardless of the direction of change of the SNIR, if the SNIR rate of change is below the standard threshold, the weight of the SNIR is set to 0 and the weight of the averaged number of failed parity checks is set to 1.

33. The device of claim 20 , wherein measuring the first error information comprises integrating a SNIR value over one radio frame and comparing its rate of change against a set of two thresholds, which includes a higher threshold being defined as a fast reaction threshold and a lower threshold being defined as a standard threshold, and the number of failed parity checks computed at said decoding iteration different from the first one is averaged over a number of LDPC blocks corresponding to a period of half the radio frame, wherein if the SNIR is decreasing and both its rate of change and the rate of change of the averaged number of failed parity checks exceed the fast reaction threshold, the weight of the SNIR value is set to 0.50 and the weight of the averaged number of failed parity checks is set to 0.50; and wherein if the SNIR is decreasing and its rate of change is below the fast reaction threshold and the rate of change of the averaged number of failed parity checks exceeds the fast reaction threshold, the weight of the averaged SNIR value is set to 0 and the weight of the averaged number of failed parity checks is set to 1; and wherein, regardless of the direction of change of the SNIR, if both the SNIR rate of change and the rate of change of the averaged number of failed parity checks are between the standard and fast reaction thresholds, the weight of the SNIR is set to 0.25 and the weight of the averaged number of failed parity checks is set to 0.75; and wherein, regardless of the direction of change of the SNIR, if the SNIR rate of change is below the standard threshold, the weight of the SNIR is set to 0 and the weight of the averaged number of failed parity checks is set to 1.

34. An adaptive modulation communication system, comprising a transmitter stage for transmitting over a communication channel data encoded through a Forward Error Correction code, said encoded data being modulated according to the communication channel conditions, wherein the system comprises a receiver stage according to the device of claim 18 .